Design and Implementation of an Artificial Vision System for Quality Control in the Labeling of Commercial Products

Large industries are continually looking for strategies and tools that help them increase their productivity, to cover larger markets in shorter times. However, on certain occasions, these types of strategies with the increase in production speed can affect the product quality, sometimes triggering the discomfort of the clients or final recipients. This is why companies with continuous production lines, in addition to seeking productive optimization strategies, nowadays seek strategies that help them maintain the quality of their products. This work presents the development of a classifier algorithm through road inspection that verifies the correct labeling of different commercial products, thus helping to maintain the final quality of the products in a production line. The algorithm has been developed in an open-source programming language with a camera of medium characteristics, to reduce commercial costs and make it a strategy for easy massification. In addition, the algorithm can be easily adapted for different types of products, that is, it is an open and undeveloped strategy for a single type of label. The algorithm has an interface that will help users and interested parties in its use. At the end of the work, an artificial vision labeling identification system is available that is capable of detecting flaws in any type of label at an estimated speed of 2 sec per product. Keywords: artificial vision, Python, OpenCV, labeling, automatic classification. Resumen Las grandes industrias continuamente andan buscando estrategias y herramientas que las ayuden a incrementar su productividad, para cubrir mercados más amplios en tiempos más cortos. En ciertas ocasiones este tipo de estrategias sacrifican la calidad de los productos por el incremento de velocidad de producción, desencadenando en algunas ocasiones el malestar de los clientes o destinatarios finales. Es así como las empresas con líneas de producción continua, a más de buscar estrategias de optimización productiva, hoy en día buscan estrategias que les ayuden a mantener la calidad de sus productos. En este trabajo se presenta el desarrollo de un algoritmo clasificador mediante inspección visual que verifica el correcto etiquetado de distintos productos comerciales, aportando así a mantener la calidad final de los productos en una línea productiva; el algoritmo se ha desarrollado en un lenguaje de programación de código abierto y con una cámara de características medias, con el fin de reducir los c